Astaneh-AISC-Blast-Resistant-Final Report-Oct 23, 2001a.pdf

7/27/2019 Astaneh-AISC-Blast-Resistant-Final Report-Oct 23, 2001a.pdf

1/44

PROGRESSIVE COLLAPSE RESISTANCEOF STEEL BUILDING FLOORS

FINAL REPORT

By

Abolhassan Astaneh-Asl

Brant Jones

Yongkuan Zhao

Ricky Hwa

With Contributions from:

David McCallen and Charles Noble

of Lawrence Livermore National Laboratory

REPORT NUMBER: UCB/CEE-STEEL-2001/03DEPARTMENT OF CIVIL AND ENVIRONMENTAL ENGINEERINGCOLLEGE OF ENGINEERINGUNIVERSITY OF CALIFORNIA AT BERKELEY

Report to the Sponsors: American Institute of Steel Construction

August 2001

This is a 99% Version of Final Report of a UC-

Berkeley research project supported by AISC. Itis submitted to AISC for its review and

comments within 30 days. The report should not

be distributed to others outside AISC. The

information cannot be used in anyway until the

report is finalized and submitted to AISC as

Final Report by Principal Investigator: A.

Astaneh-Asl.

October 23, 2001


2/44


3/44

FLOOR CATENARY ACTION TO PREVENT

PROGRESSIVE COLLAPSE OF STEELSTRUCTURES

FINAL REPORT

Abolhassan Astaneh-Asl, Ph.D., P.E.

Professor and Principal InvestigatorDepartment of Civil and Environmental Engineering

University of California, Davis HallBerkeley CA 94720-1710

Brant Jones

Master of Science Graduate Student Research Assistant

Yongkuan Zhao

Associate Professor at Zhaoquing University, Peoples Republic of China andVisiting Scholar, University of California at Berkeley (2001)

Ricky Hwa

Undergraduate Student Research Assistant

With Contributions from:

David B. McCallen and Charles Noble

Of the Lawrence Livermore National Laboratory

Report to the American Institute of Steel Construction

REPORT NUMBER: UCB/CEE-STEEL-01/03DEPARTMENT OF CIVIL AND ENVIRONMENTAL ENGINEERING

COLLEGE OF ENGINEERINGUNIVERSITY OF CALIFORNIA AT BERKELEY

August 2001


4/44

Technical Report Documentation Page1 Report No

UCB/CEE-STEEL-01/03

2. Government Accession No. 3. Recipients Catalog N

4. Title and Subtitle

FLOOR CATENARY ACTION TO PREVENT

PROGRESSIVE COLLAPSE OF STEEL STRUCTURES

5. Report Date

August 2001

6. Performing Organization Code

7. Authors

A. Astaneh-Asl, Brant Jones, Yongkuan Zhao, Ricky Hwa,8. Performing Organization Report No.

UCB/CEE-STEEL-01/039. Performing Organization Name and Address

Department of Civil and Environmental Engineering10. Work Unit No. (TRAIS)

University of California, Berkeley, California 94720-1710

Fax: A. Astaneh: (510) 643-5258, e-mail:[email protected]

11. Contract or Grant No.

Contract Number:

12. Sponsoring Agency Name and Address

American Institute of Steel Construction13. type of Report and Period Covered

Final

15. Supplementary NotesThe research reported herein was done in the Department of Civil and EnvironmentalEngineering of University of California, Berkeley and was supported by the American Institute ofSteel Construction. The construction of full scale specimen used in these studies was funded by the

General Services Administration of U.S. Federal Government in a sister project. The opinionsexpressed in this reports are those of the authors and do not necessarily reflect the views of

University of California, Berkeley or the Sponsor.

16. Abstract

This report presents the background, test set-up, specimens, test procedures and the test results of

column-drop tests of a one story steel structure. The specimen was a 60ft by 20ft one story typicalsteel structure with steel deck and concrete slab floor and wide flange beams and columns. The

connections were either standard shear tab or bolted seat angle under bottom flange and a boltedsingle angle on one side of web.

The main objectives of these studies were to explore the strength of a typical steel structure andfloor system to resist progressive collapse in the event of removal of a column by a blast and to

establish failure modes. An added objective was using the test results to provide the AISC withdesign-oriented information on what is the potential of existing typical steel structures to resistprogressive collapse and what are the possible research needs in this particular field.

The tests indicated that after removal of the middle perimeter column, due to catenary action of steeldeck and girders, the design dead load and live load of the floor could be resisted and the floor is not

expected to collapse in the event of removal of one such column.

17. Key Words

Steel Structures, Blast-resistant Design,Progressive Collapse Prevention, FloorDesign, Steel Deck, Bolted Connections,Testing.

18. Distribution Statement

No restriction. This Document is available to thepublic through the Department of Civil andEnvironmental Engineering, University ofCalifornia, Berkeley, CA 94720-1710Contact: [email protected]

19. Security Classif. (of this Report)

Unclassified20. Security Classif. (of this page)

Unclassified21. No. of Pages 22. Price

$40.00+Applicable Sales Tax


5/44

ABSTRACT

This report presents the background, test set-up, specimens, test procedures and the testresults of column-drop tests of a one story steel structure. The specimen was a 60ft by 20ft

one story typical steel structure with steel deck and concrete slab floor and wide flange beamsand columns. The connections were either standard shear tab or bolted seat angle under bottom

flange and a bolted single angle on one side of web.The main objectives of these studies were to explore the strength of a typical steel

structure and floor system to resist progressive collapse in the event of removal of a column by ablast and to establish failure modes. An added objective was using the test results to provide the

AISC with design-oriented information on what is the potential of existing typical steelstructures to resist progressive collapse and what are the possible research needs in this particular

field.The tests indicated that after removal of the middle perimeter column, due to catenary

action of steel deck and girders, the design dead load and live load of the floor could be resisted

and the floor is not expected to collapse in the event of removal of one such column.


6/44

ACKNOWLEDGMENTS

The study reported here was supported by the American Institute of Steel Construction.

The authors would like to express their appreciation for the support as well as technical advice

and input received from the AISC engineers. Particularly, the support and technical input from

Dr. Nestor Iwankiw, Dr. Sergio Zoruba and Mr. Tom Schlafly of the AISC were very useful and

essential to success of this project and are sincerely appreciated.

The tests reported here were conducted in the Department of Civil and Environmental

Engineering of University of California at Berkeley. The contributions of Department staff

particularly William Mac Cracken, Dr. Lev Stepanov, Chris Moy, Richard Parsons, Douglas

Zulaica and Mark Troxler were very valuable in conducting the tests as well as in data

acquisition and processing. Without their dedication and professional work, this project would

not have been possible. Roger Jung, Erik Madsen and Qiuhong Zhao , graduate student research

assistants participated in various aspects of the project. Wendy Wang and Sunil Tellis ,

undergraduate student research assistants, were also involved in preparing the specimen,

conducting material tests and other aspects of this large scale testing. The contributions of these

students were very valuable, acknowledged and sincerely appreciated.

The specimen used in these studies was constructed for a sister project that was funded

by the General Services Administration (GSA) of U.S. Federal Government. Herrick

Corporation of Stockton California fabricated the test specimen with diligence, on time and on

budget and according to drawings and specifications. The efforts of Jamie Winens, the project

engineer with Herrick Corporation are sincerely appreciated.

The testes reported herein were primarily conducted by second author, Brant Jones, as a

CE-299 Independent Research and during a summer spent on the campus of the University of

California Berkeley as part of a Professional Development Program, a privilege for which he is

grateful to the Defense Nuclear Facilities Safety Board.


7/44

TABLE OF CONTENTS

Report Title Page.................................................................................... 1

Report Summary Form . 2Abstract ......................................................................................................... 3Acknowledgments ...................................................................................... 4

Table of Contents ....................................................................................... 5

1. Introduction.......................................................................... 6

2. Experimental Research......................................................................... 9

3. Observed Behavior and Test Results...................................... 23

4. Conclusions and Recommendations................................... 36

Appendix A. Shop Drawings of Test Specimen

Appendix B. Complete Test Data on Specimen


8/44


9/44

CHAPTER ONE

Introduction

1.1. Introduction

In recent years, a number of tragic terrorist attacks, particularly, in the U.S. soil, have

resulted in a number of initiatives to study resistance of structures to blasts. In addition, a number

of research projects have been undertaken or are underway to develop mechanisms and systems

that can reduce the hazard of such attacks. The main aim of these efforts is to protect life safety

of the building occupants and those who are around the building and can be injured by collapsing

structure and falling debris.

One of the main areas of research and development in this field is progressive collapse

prevention. From structural engineering and construction point of view, of course one can design

a building that can withstand a terrorist bomb attack with minimal or no damage. This has been

done for years and continues to be done for militarily sensitive and other critical buildings that

are necessary to be functional and occupied even after an explosion attack on them. Of course,

designing such a highly protected building requires significant amount of funding as well as

resources. In addition, to achieve the objective of minimal damage, the designers may end up

sacrificing the exterior aesthetics and in some cases the internal functionality of the building.

Although in case of military installations, the high cost and bunker-like look of a building can be

justified, for civilian office buildings, such high costs cannot be afforded and the loss of aesthetic

may not be always acceptable. This is because of very low probability of any given civilian

building being attacked by terrorists.

In order to make a civilian building structure blast-resistant, first, one has to define the

performance criteria for blast resistance. The performance criteria should be specific enough to

provide structural engineers with clear parameters such that they can design the structures to

achieve the expected performance. The following section discusses Performance Criteria in more

detail.


10/44

1.2. Blast-Resistance Performance Criteria

The blast performance criteria should be established in a way that the society as a whole

is satisfied with the target performance level and can afford the associated cost. Currently there

are no formal blast performance criteria for civilian buildings. It seems that a criteria that may

be acceptable and affordable by the society at this time and with present low probability of

terrorist attacks is to ensure that the loss of life and severe injuries are avoided and partial or full

collapse are prevented. Following this proposed performance criteria, damage to the building in a

terrorist attack is accepted as long as such damage does not result in death or severe injuries or

collapse.

There are several reasons for death and injuries in a blast attack. Death and injury can

occur due to direct impact of explosion, heat, fire, inhaling smoke and hazardous gases

eminating from the explosion or building contents, flying debris or building contents, falling

debris or building contents, shrapnels, air impact and other related causes such as heart attack.

Collapse of a structure and the resulting pancaking of floors has been the main cause of

deaths and injuries in recent blast attacks on buildings. It is clear that by preventing collapse,

especially by preventing catastrophic collapse of floors, one can save many lives and be able to

evacuate the occupants to safety. Therefore, making civilian buildings blast-resistant can be

translated to making sure that partial and full collapse of the structure of these buildings are

prevented.

Some structures have highly indeterminate structural systems with more than one load

paths to resist the applied loads. In these highly indeterminate systems with many redundant

members, it is possible to eliminate a number of structural elements yet not to be able to collapse

the structure or the floors. Obviously, if such systems are identified, their use will result in

significant saving in making building structures no to collapse during or after an explosion

attack.. Such characteristics denoted as progressive collapse resistance was the main theme of

this research project.

1.3. Objectives

The tests reported herein were conducted to collect data on progressive collapse

resistance of typical steel structure with steel deck concrete slab floor supported on steel framing

with simple connections.


11/44

1.4. Scope and Content of the Report

The focus of this report is the investigation of conventionally designed buildings in their

ability to resist progressive collapse. The discussion will be limited to the observed behavior and

test data resulting from experimental research of a full-scale floor system of a typical commercial

office building. Conclusions and recommendations will focus on the overall failure mechanisms.

Some recommendations are made for future research needs.


12/44

CHAPTER TWO

Experimental Research

2.1. Introduction

Early in 2001, a series of tests were conducted at University of California at Berkeley on a

full scale one-story steel structure to study a cable-based mechanism that could prevent

progressive collapse of the steel structures. The project funded by the General Services

Administration, was successfully completed and the results were reported in Reference (

Astaneh-Asl, Madsen and Jung, 2001). The tests indicated that by placing steel cables inside the

floor slab along the column lines, one can prevent collapse of the floor if a supporting column is

removed by blast forces during a terrorist attack. The specimen used in these tests was a 60ft by

20 ft floor supported by a steel structure with 8 columns. The one-story structure, shown in

Figure 2.1, was a realistic representative of typical steel structure.

Figure 2.1. Test Specimen

PLAN

Cables W14x61

All N-SBeams

W21x44Deck Ribs

Anchor Stand

1

3 1/2" conc over 3"

Normal wgt.

20GA WJ deck

with wwf 6x6

18x318x35

2

18-0

A

NW18x35 W18x3

W18x35W18x35

20 2010

10-0 10-020-0 20-0

Drawing by A. Astaneh-AslELEVATION

Cables


13/44

After completion of the above-mentioned GSA tests, which were conducted on one side

of specimen (the North side), the other side of specimen was almost undamaged with minor

cracking in the concrete slab. It was decided to take advantage of availability of the specimen

and to conduct similar tests on the south side of specimen, where there was no cable, see Figure

2.1. This report is focused on the tests conducted on the South side of specimen where there

was no cable. The reader is referred to Reference (Astaneh-Asl, Madsen and Jung, 2001) for

information on the tests conducted on the North side of specimen where steel cables placed

within the floor slab.

2.2. Objectives of the Research and Testing

The main objectives of the research and testing reported herein were:

a. To investigate potential of typical steel floors to develop catrenary action in the event of

removal of a supporting column.

b. To establish potential of typical steel structures to resist progressive collapse by

redistributing column loads through Catenary action of floors.

c. To develop a better understanding of how connections of steel structures perform when

subjected to large deformations and large Catenary forces.

d. To develop recommendations for future research as well as tentative suggestions for

blast-resistant design of typical steel structures.

2.3. Design of Specimen

The structural design of test specimen was done by SWMB engineers. The Principal

Investigator at the University of California Berkeley (UCB) was involved in the design

process, ensuring that the specimen was designed within the testing limitations and would fit

properly the test set-up and the laboratory reaction floor. The specimen consists of steel frames

with simple shear connections and a concrete slab cast on a steel deck. The structure was

designed to satisfy the applicable code requirements for a typical structure in Seattle,

Washington.

The specimen, shown in Figure 2.1, had four bays in the longitudinal direction totaling 60

feet and one bay in the transverse direction with a width of 18 feet. The exterior bays were both

10 feet long while the interior bays were 20 feet. The concrete slab was 6 feet above the


14/44

laboratory floor level. Design drawings as well as shop drawings of the specimen are provided

in Appendix A.

The design produced W14x61 steel columns, W18x35 beams in the longitudinal

direction, and W21x44 beams in the transverse direction at 10 feet intervals. Figures 2.2 tand 2.3

show framing plan and elevation view of the test specimen.

Figure 2.2 Framing Plan of Test Specimen

Figure 2.2 Framing Plan of Test Specimen

Figure 2.4 shows a cross section of test specimen. The right side frame was tested in this

program.


15/44

Figure 2.3. Cross Section of Test Specimen

The steel deck was a typical 20 gage deck connected to the framing with puddle welds

and with shear studs. Figure 2.4 shows a general view and close up of the deck.

Figure 2.4. A view of Stel Deck and Close-up of Deck and Shear Studs

Figure 2.5. shows plan view of the shear studs. Placed on top flange of beams and

girders.

Figure 2.5. Plan of Shear Studs


16/44

The floor slab in the test specimen was a 6.5 inch maximum thickness slab. The North

side which had cables in it had reinforcement but the South side, the subject of tests reported

here did not have any additional rebars other than the standard wire mesh as shown in Figure 2.1.

Figure 2.6. shows plan of reinforcement and details. Figure 2.7 shows a plan views of slab

around the columns.

Figure 2.6. Plan of Reinforcement and Details

Figure 2.7 Close-up View of Slab in the Column Areas


17/44

Figure 2.8 shows The connections in the specimens were all either standard shear tabs or

bolted seat angles plus a bolted single angle on the web. Figure 2.7 shows details of the

connections.

Figure 2.8. Beam-to-Column (left) and Beam-to Girder (right) Connections


18/44

2.3. Construction of Specimen

The steel members and connections of the test specimen were fabricated by Herrick

Corporation and delivered to Davis Hall laboratory of UC-Berkeley. The steel framing was

erected inside the lab by UC-Berkeley machinists and welders. o be added later. After installing

the deck, shear studs were installed, wire mesh and rebars added and concrete slab cast. The

construction process was similar to constructing a typical steel structure. Figure 2.9 shows

various stages of construction.

Figure 2.9. Construction of Test Structure


19/44

2.4. Material of Test Structure

The steel mill certificates are given in Appendix A. The beams, columns, angles, and

shear tabs were specified as A36. The concrete in the floor slab was specified as normal

weight concrete with fc = 4000 psi.. The slump of the concrete was measured to be 4

inches, Figure 2.9. Concrete tests confirmed a compressive strength greater than 4000 psi

after 21 days. The concrete tests were performed at UC-Berkeley material lab. The results of

tests are shown in Figure 2.10.

Figure 2.9. Slump Test and Cylinder Tests on Concrete Used in Slab

Figure 2.10. Results of Concrete Compression Tests


20/44

The shear tabs and angles in the connections were specified as A36. All bolts were

A325, 7/8-inch diameter. The lengths were governed by the widths of the connecting

material, with either 2 inch or 3-inch bolts providing the required length. The welds were

all performed in the shop and were performed by Herrick Corporation.

Verco Structural Steel Decking provided the Type W3 Formlok, Gauge 20 steel deck.

The inch diameter, 4 inch long shear studs used in the test structure were supplied and

installed by Nelson Studs. They were placed at 8 inch centers along the longitudinal beams

and 1-0 centers along the transverse beams as shown in Figure 2.5. W1.4xW1.4 flat

sheet reinforcement with 6 inch by 6-inch openings was utilized within the concrete floor

slab.

2.5. Instrumentation of Specimen

Displacement transducers and strain gages were used at various location on the specimen.

Details of instrumentation plans are given in Appendix B along with the complete data obtained

from the instruments.

Displacement transducers were positioned at various locations on the structure to

measure the movement of the structure throughout the tests. Strain gauges were installed on the

structure to monitor the strain at locations of interest.

Nineteen displacement transducers were positioned on and around the test structure.

Specific locations of the transducers are included in Figures 3.11,3.12 and 3.13. Transducers

located vertically around the specimen measured the drop of the floor at those locations while

horizontal transducers measured the translation. Pairs of displacement transducers located on the

beams connecting into the test column were used to measure the rotational demands of the

connections. When possible, the transducers were aligned parallel to the expected direction of

displacement.

Six transducers were positioned horizontally around the outside of the structure to

monitor the in-plane translation of the specimen. The transducers were secured to rigid supports

independent of the test specimen. Connecting wires were secured to the structure and tied to the

transducer. The horizontally placed transducers were labeled Displacement Transducer (DT) 1,

3, 14, 16, 20, and 22.


21/44

A pair of transducers was placed on the ends of each of the three beams framing into the test

column. An additional pair was placed at the opposite end of the transverse beam. The

transducers were aligned parallel to the webs of the beams, two inches from the outermost

surface of the closest flange. A threaded rod was welded to the web of the beam onto which

each transducer could be attached. A connecting wire connected the transducer to the test

column. The eight rotational transducers were labeled DT 6 through 13.

Three displacement transducers were positioned along the test column line to measure the

vertical movement of the frame. One transducer was positioned at the test column while two

other transducers were placed near the adjacent columns. The three transducers were labeled DT

2, 4, and 15.

Two additional displacement transducers were attached to the test column as can be seen

in Figure X. The transducers were labeled DT 17 and 19. The test column was not only

expected to displace vertically, but also to rotate as it displaced. For this reason, it was not

possible to orient the transducers in a way to reduce the measurement to one dimension. Initial

connecting wire lengths were recorded and are included in the figure.

Displacement transducers with labels numbered 5, 18, and 21 were not utilized in the

testing. The transducers were labeled in a fashion so that there would be one-to-one

correspondence to the transducers used in previous experiments of the specimen.

Strain gauges were installed to monitor the strain at locations of interest. Specific

locations of the gauges are included in Appendix B. Four 45 rosettes were placed in a

symmetrical fashion on the underside of the steel deck. Three linear gauges previously attached

to an angle were monitored. The particular angle connected the test column to the longitudinal

beam directly east of the test column. Also, four linear strain gauges were placed underneath the

steel deck in a single line, 30 inches east of the test column. And seven additional gauges were

placed on the web and flanges of the longitudinal beam directly east of the column, also 30

inches from the test column.

The specimen was carefully prepared before the strain gauges were bonded to the surface.

Rust and mill-scale were removed from the steel beam and angle by grinding the surface with a

stone-based grinding wheel. The surface was then progressively smoothed by using a cycle of

sanding wheels beginning with a 36-grit wheel and ending with a 120-grit wheel. This process

removed visible defects. Next, the surfaces were sanded by hand using 240-grit sandpaper


22/44

followed by 360-grit sandpaper. The process of placing the gauges on the steel deck began at

this hand-sanding step. The locations were then ready to be chemically treated.

Before chemically treating the locations, the exact location of the gauge was scored into

the steel material. A water based acid surface cleaner applied with the 360-grit sandpaper was

then used to cleanse the surface. Next, a degreaser was applied and wiped clean to remove any

acid that remained on the surface. Wiping the surface with the acid surface cleaner immediately

followed by a swipe of an alkaline surface cleaner then neutralized the surface. The water based

alkaline surface cleaner contained ammonia water whereas the acid cleaner contained phosphoric

acid. At this point, the surface could be visually inspected, with colored streaks revealing any

contamination that required a subsequent neutralization step. Finally, the strain gauges were

secured to the surface using a high-strength glue, cyanoacrylate adhesive, and protected with a

layer of duct tape.

Four 45o strain rosettes were placed on the underside of the steel deck in the bays

adjacent to the displaced column. These gauges were symmetrically placed, with two residing

on the west side of the test column, and two on the east side. During installation and inspection,

it was determined that both longitudinal (east-west) gauge wires in the bay immediately west of

the test column were damaged and did not provide the necessary resistance. For this reason, they

were not attached to the data recording system. No problems were encountered with the adjacent

strain rosettes, placed in the bay immediately east of the test column. The ten functional gauges

were labeled Strain Gauge (SG) 1 through 10.

Three linear gauges were placed on an angle connecting the test column to one of the

longitudinal beams during testing of the opposite side (See Appendix B for location of gages).

The gauges were once again monitored during testing. Significant and visibly noticeable

yielding of the angle occurred during the prior testing. The three linear strain gauges were

labeled SG 11 through 13.

Four additional linear strain gauges were placed beneath the steel deck, 30 inches east of

the test column. The gauges provided data on the variation of stress in the steel deck. The four

gauges were labeled SG 14 through 17.

The seven gauges that were placed on the longitudinal beam directly east of the column

provided data that could be used to calculate the force in the beam throughout the tests. Two

gauges were place on the bottom of the top flange and the top of the bottom flange, one on either


23/44

side of the web. Three additional gauges were placed on the web of the beam facing the outside

of the structure. The seven gauges were all located 30 inches east of the test column and labeled

SG 18 through 24.

2.6. Data Acquisition

The displacement transducers and strain gauges were attached to six data acquisition

boxes. Each box was capable of monitoring eight channels. The instruments were attached such

that only one of the two types of instruments led into each of the boxes. The acquisition boxes

were subsequently connected to a Megadak data recording system. With a channel capacity of

56 and the capability to record 50 data points per second, the system provided sufficient data

recording capabilities.

A second computer controlled the actuators during the experiments. Not only was the

computer capable of recording the force and displacement in each actuator, but also a small

number of additional channels. For this reason, several of the key instruments were also

connected to the second computer to serve as back-up system. The data was recorded and

processed at a later date.

2.7. Problems Encountered in Instrumentation and Data Acquisition

Examination of the data revealed that the displacement transducers (DTs) numbered 4

and 19 both produced unexpected output. DT 4 was positioned just outside the test column and

measured the vertical drop in the test column. The processed data plots of DT 4 reveal changes

in the direction of movement. With the actuators displacing the column at a constant rate, the

processed data was clearly unreliable. The transducer may have been incorrectly attached to the

data acquisition system.

DT 19 was attached to the structure near the top of the test column, however the layout of

the lab required that the transducer be secured to piping located on the ceiling of an adjoining

sub-room. When securing the transducer to the overhead piping, it was believed to be sturdy

enough to resist the small force needed to extend the wire. However, the processed data clearly

shows that the transducer was displaced approximately 0.4 inches by this small force. This is

evident because once the displacement of the test column was completed, the transducer shows a


24/44

shortening of the connected distance by about 0.4 inches. Follow-up inspection of the pipes to

which it was attached confirmed the likelihood of this transducer movement.

2.8. Test Procedures

Prior to the testing date, all instrumentation was installed, calibrated, and tested for

proper operation. This included a check of the strain gauge resistances. The test consisted of

three constant velocity loadings. The column was displaced at a rate of 0.25 inches/second to

three different displacement levels. The column was displaced 19, 24, and 35 inches downward

for the first, second, and third subtests, respectively. After each subtest, the bottom of the test

column was returned to its original height, 36 inches above the base level. The final

displacement of each subtest was governed by the tests that were performed on the opposite side

of the structure that contained the catenary cables. By performing the same tests on each side,

one can evaluate the contribution of the catenary cables in resisting the increased load.

Figure 2.11. The Two Actuators and Loading Mechanism Used to Push the Column Down

The specimen was tested with the use of two 120,000-pound hydraulic-based testing

machines located on the second floor of Davis Hall on the main campus of UCB. The tests were

performed on June 14, 2001. A photograph of the test specimen can be seen in Figure 2.11

Actuators


25/44

showing the loading actuators that were used to push the floor down. In order to remove the 36

inch support column, the actuators were raised a slight amount, leaving a gap slightly greater

than 36 inch between the bottom of the column and the base of the floor. Once the column

support was removed, the specimen was ready to be tested at the prescribed column

displacements.


26/44

CHAPTER THREE

Observed Behavior and Test Results

This chapter summarizes the test results and observed behavior of the test specimen. A

complete set of test results and data recorded during the tests is provided in Appendix B.

3.1. Test 2A (0 to 19-inch Displacement)The first test was conducted at 2:50 PM on June 14

th, 2001. The test column was

displaced 19 inches in this test, identified as Test 2A. The force time history and the force

displacement plots can be seen in Figure 3.1. The total force was calculated as the sum of both

actuator forces.

Figure 3.12 Force Time History and Force versus Displacement for 19-inch Drop Test


27/44

As the column displacement exceeded 14 inches, failure occurred of the top two bolts on

the lower leg of the seat angle that holds the longitudinal beams. This is identified as spike A

in both the force time history and force displacement plots in Figure 3.1. The end of the bolts,

including the nuts, violently shot through the air while the head of the bolts remained in place. A

total of four bolts held the angle to the displacing column, leaving only two bolts to resist the

vertical load of the beams. As the bolts were failing, the seated angle itself was undergoing local

yielding and deformation. See Figure 3.2 for a photograph of the local damage.

Figure 3.2. Failure of Top Two Bolts in the Seat Angle

As the displacement approached 15 inches, at the other end of one of the longitudinal

beams, west of the displaced column, the two bolts that secured the beam to the top of the seated

angle also failed and popped (Figure 3.3). Spike B in Figure 3.1 clearly shows the loss of

strength as a result of the bolt failures. The shear tab that supports the transverse beam also

underwent local yielding.

As the displacement approached 17 inches, slippage at some point within the structure

occurred. This is evident by examining spike C in Figure X. Although the load carrying

capacity of the structure abruptly decreased at that point, the load gradually increased to the same

value as before the spike. A failure would result in an unrecoverable decreased capacity,

whereas slippage would only result in temporary decreased load carrying capacity.


28/44

Figure 3.3. Failure of the Two Bolts on the Seat Angle

The concrete slab separated from certain columns as the displacement increased. At the

final displacement of 19 inches, a gap of approximately 1 inch had opened between the slab and

column B2. Photographs of the local damage surrounding columns B2, C2 and D2 are included

as Figures 3.4(a), 3.4(b) and 3.4(c) respectively. Figure 3.4(d) shows close-up of the cracks

around column B2.

Figure 3.4. Cracking of the Concrete Slab at Column B2 at the End of 19-inch Drop Test

a


29/44

After the column reached a displacement of 19 inches, the displacement was held

constant for approximately one minute before returning to a displacement of 17 inches. The total

sustained force, including dead load, at the final displacement of 19 inches was 54.0 kips. While

the structure was displaced 17 inches, the research team inspected and made notes of the relevant

deformations. Approximately 10 minutes later, the column was returned to the original height,

corresponding to zero displacement.

Next, the two bolts that failed at the loaded column were replaced. Due to the seated

angle deformation noted above, a inch gap remained along the shaft of the bolt after the nut

was tightened.

Figure 3.5. A Sketch and Views of Specimen at the End of 19-inch Drop Test

A sketch of specimen at the end of 19-inch drop test is shown in Figure 3.5. The most

notable behavior observed in the 19-inch displacement was the failure of two sets of bolts. The

first to fail, located at the seated connection of the displacing column, revealed the limited

rotational flexibility of the beam-to-column connections. The bolts failed when the longitudinal

beam rotated 3.2relative to the displaced column. Immediately following, the bolts failed at the

other end of the beam directly west of the displacing column. This second set of bolts connected

the longitudinal beam to the seated connection. Although these bolts do not provide any direct

vertical load carrying capability, the catenary action of the displacing frame is lost.

A maximum micro-strain of 1450 was measured on the steel deck during the

displacement. The corresponding maximum stress was calculated to be 42.1 ksi. Four linear

strain gauges were placed on the deck, 30 inches from the centerline of the structure. The

highest strain was measured at the innermost strain gauge, with strains generally decreasing as

location approached the edge of the structure. The notable difference is the second innermost


30/44

strain gauge, which was placed about an inch from the end of a steel deck segment. Being so

close to a free edge, the deck was not able to develop the relatively large strains of the adjacent

strain gauge locations.

The outermost strain gauge recorded the lowest strain value. Post-experiment inspection

of the structure revealed significant crushing of the rib on which the gauge was located. The

local crushing was concentrated above the transverse beam connecting the two center columns.

It is believed that this local crushing was responsible for relieving the stress in the rib containing

the outermost strain gauge. As a result, the strain gauge, located over two feet away from the

local crushing, recorded low levels of strain. Also, the high stiffness of the longitudinal beam

located just 13 inches away may have reduced the strain in the outermost gauge.3.2.

Test 2C (0 to 24-inch Displacement)

The second test was conducted at 3:15 PM on June 14th, 2001. The test column was

displaced 24 inches in this test, identified as Test 2C. The force time history and the force

displacement plots can be seen in Figure 3.6. The total load in Figure 3.5 was zeroed at the

beginning of the test in the data processing stage.

Figure 3.6. Force Time History and Force-Displacement Plot of 24-inch Drop Test.


31/44

As the column reached a displacement of approximately 19 inches, the two bolts that held

the longitudinal beam, east of the loaded column, on the seated connection at column C2 failed

(Figure 3.7). The dramatic drop in load is evident as spike D in Figure 3.6. Local yielding of

the seated connection could also be observed. Also, cracking and tearing of the connecting angle

between the loaded column and the longitudinal beam directly east had begun.

As a result of the two bolt failures, large eccentricities were created as the east actuator

shot back to zero load while the west actuator was still providing 20 kips of load to provide the

constant displacement. At a displacement of 24 inches, the west actuator was applying

approximately 45 kips of downward force while the east actuator was applying approximately 10

kips of upward force.

Figure 3.7. The Two Bolts on The Seat Angle Were no Longer Present

The steel deck suffered noticeable damage. Major deformations were noted along the

column line of the displaced column, with concentrated damage occurring near the beam to

column connections. The steel deck was ripped near columns B2 and D2, while bending was

widespread.


constant for approximately one and a half minutes before returning to a displacement of 20


32/44

inches. The total sustained force, including dead load, at the final displacement of 24 inches was

62.8 kips. While the structure was displaced 20 inches, the research team inspected and made

notes of the relevant deformations. Approximately 10 minutes later, the column was returned to

the original height, corresponding to zero displacement.

While attempting to replace the same top two bolts on the seated angle connection that

were replaced before, it was noticed that the bolts did not fail, but underwent large deformations.

The bolts were removed, but could not be replaced with new bolts due to the large deformations.

Figure 3.8. The Two Replaced Bolts on the Top Half of the Bottom Leg of the Seat Angle

Remained in Place.

A sketch of specimen at the end of 24-inch drop test is shown in Figure 3.9. Only one

failure occurred during the column displacement to 24 inches. The pair of bolts failed that

secured the longitudinal beam directly east of the displaced column to the seated connection.

Unlike the beam west of the column, this failure occurred at the end connecting into the test

column. Both failures left the bottom flanges of the beams free to move at one end. As a result

of the asymmetrical failure, the structure was no longer behaving in a symmetrical manner. The

east actuator was actually applying a 10 kip upward force to the column even though the column


33/44

was being displaced downward. Strengthening the bolts to eliminate their failure should produce

a more balanced, well-behaved failure mechanism.



34/44

The force-displacement plot (Figure 3.6) shows a gradual stiffening of the structure once

the displacement passes 5 inches. This results from the replacement of the two bolts in the

seated connection at the displaced column. The inch gap that remained after the replacement

affected the behavior of the structure. As the bolts adjusted to the gap and engaged themselves,

the structure became stiffer.

A maximum micro-strain of 1560 was measured during the displacement. The

corresponding maximum stress was calculated to be 45.2 ksi. See Appendix B for the maximum

calculated stresses at the instrumentation locations. The strains for this second subtest include

any residual strains that existed within the gauges after the column was returned to a position of

zero displacement. As before, the innermost of the four linear strain gauges placed on the deck

recorded the highest strain. The strain generally decreased as the edge of the structure was

approached with the exceptions noted above.

3.3. Test 2D (0 to 35-inch Displacement)

The third test was conducted at 4:10 PM on June 14th, 2001. The test column was

displaced 35 inches in this test, identified as Test 2D. The force time history and the force

displacement plots can be seen in Figure 3.10. The total load in Figure 3.8 was zeroed at the

beginning of the test in the data processing stage.

In addition to the above noted behavior, as the column displacement approached 26

inches, complete failure occurred of the connecting angle between the loaded column and the

longitudinal beam directly east (Figure 3.10). Spike F in Figure 3.8 shows the sudden drop in

load after the angle failure. The preceding spike, labeled E, is evidence of slippage within the

structure, possibly within the angle before complete failure.

After the angle had failed, the concrete slab was the primary element transferring the

applied force to the longitudinal beam directly east of the displacing column. Unable to maintain

this force transfer, the slab failed in shear along the transverse beam connecting the two central

columns. This is represented by spike G in Figure 3.10.


35/44

Figure 3.10. Force time history and force-displacement plot of Test 2D.

Figure 3.11. Complete Failure of the Shear Angle on the Web of Column


36/44

The concrete slab and steel deck suffered great damage during the 35-inch displacement.

The steel deck continued to tear, opening up gaps both within and between deck segments. The

damage was concentrated along the column line of the tested column. There was a separation of

at least one inch between deck segments along the transverse beam connecting the two center

columns. Near column D2, the deck tore, opening up a hole that measured over three inches

between the steel deck and concrete slab.


constant for approximately 2 minutes before returning to a position of approximately zero

applied force. The total sustained force, including dead load, at the final displacement of 35

inches was 43.5 kips. The structure remained in the position of zero applied force,

corresponding to a displacement of approximately 27 inches, for several days. While the

structure was displaced 27 inches, the research team inspected and made notes of the relevant

deformations. The following week, the column was returned to the original height,

corresponding to zero displacement.



37/44

In addition to the above noted failures, two additional component failures occurred

during the 35-inch displacement. First, failure occurred of the connecting angle between the test

column and the longitudinal beam directly east. The concrete slab was left as the only remaining

component to transfer the applied force of the actuators to the east side of the structure. Unable

to provide this force transfer, the concrete slab failed in shear.

A maximum micro-strain of 1340 was measured during the displacement. The

corresponding maximum stress was calculated to be 38.9 ksi. See Figure X for the maximum

calculated stresses at the instrumentation locations. The gauge readings include any residual

stress that remained in the deck after the column was returned to zero displacement following the

previous test. Once again, the innermost of the four linear strain gauges placed on the deck

recorded the highest strain. The strain generally decreased as the edge of the structure was

approached with the exceptions noted above.

3.4. Summary of the Behavior and Results

Of the three subtests, the 24-inch displacement sustained the highest load of 62.8 kips.

The 19-inch and 35-inch test sustained 54.0 and 43.5 kips of loading, respectively. Figure X

provides a graph of the force-displacement relation of the three tests. As can be seen in the

figure, a maximum load is reached at a displacement somewhere between 19 and 35 inches.

The load-displacement plots of all three tests are included on one plot in Figure 3.13.

The proximity of all five component failures can easily be seen, beginning with the initial bolt

failures and ending with the complete transverse cracking of the floor slab. Also, the stiffness

degradation of the system is evident as the same load results greater displacements in the

following tests. The capacity of the structure was reached just before failure of the pair of bolts

that secured the east beam onto the seated connection at the test column. This failure, occurring

during Test 2c, resulted in the largest drop in applied load.


38/44

Figure 3.13. Force-Displacement Plot of the Three Tests.

0

10

20

30

40

50

60

70

0 5 10 15 20 25 30 35 40

Displacement (inches)

Force

(kips)


39/44

CHAPTER FOUR

Conclusions and Recommendations

4.1. Conclusions

The following conclusions were reached by observing the behavior of the test specimenand by studying test data obtained.

1. The ultimate capacity of the structure following a loss of column is limited by the beam-

to-column connections at the lost column. The catenary action of the beams and steel

deck appear to be adequate to prevent progressive collapse provided that connection bolts

do not prematurely fail.

2. The steel deck is effective in redistributing the increased load resulting from the loss of a

column. Maximum stress readings during the tests were mostly between 5 and 15 ksi,

however, localized small areas of deck yielded. Failure of the steel deck through ripping

or tearing was isolated and did not limit the strength of the system.

3. The combined Catenary action of floor steel deck and the simply-supported girders was

able to prevent collapse of column with a load of about 63 kips in the column. This load

corresponds to about 300 kips per square feet of tributary area of the floor. Considering a

reduction factor of 0.5 due to impact, the floor load that was carried by Catenary action

could be established as about 150 pound per square feet of tributary area.


40/44

4. This one test only established the potential of typical steel structures to resist progressive

collapse in the event of sudden removal of a column. Further research is needed to

establish the parameters that affect this resistance and to develop appropriate design

guidelines to take advantage of this phenomenon in preventing collapse of structures.

5.2. Recommendations

The following recommendations are based on the observed behavior of the test specimen

and by studying test data obtained. Each recommendation is directly related to one of the

observed failure mechanisms.

1. Connection bolts should be strengthened to improve structure behavior in a loss of

column scenario. The bolt failures observed in the experiments led to greater specimen

deformations and resulted in failure of a beam-to-column angle connection. High-

strength bolts and/or larger sized bolts would improve the behavior of the system with

little additional cost.

2. Placing reinforcement bars within the concrete slab could postpone and possibly

eliminate complete cracking of the slab. The bars could be placed along the column lines

of any frames susceptible to blast loads. The additional steel cross section should prevent

complete shear fracturing of the floor slab.

3. The connections should be designed for combined effects of bending and axial load due

to Catenary forces. Guidelines for design of such connections can be developed upon

further research, experimental or analytical, and better understanding of magnitude of

forces and rotations involved. The first author will continue efforts in that direction.


41/44

BIBLIOGRAPHY

ASCE, (1996). The Oklahoma City Bombing: Improving Building Performance ThroughMulti-Hazard Mitigation. FEMA 277. Federal Emergency Management Agency MitigationDirectorate. Reston, VA.

Auster, B. (2000). Striking Back at Terrorism.Prism, American Society of engineering

Education, Feb., 17-21

Baker, J.F., Lax, D. & Williams, E.L. (1948). The Civil Engineer at War. The Institution ofCivil Engineers. London, England.

Brokaw, J.T., Smith, J.L. & Swatzell, S.R. (1998). Progressive Collapse Analysis and Design

Guidance: Summary Report Documenting the Literature Review Findings. Volume I. AppliedResearch Associates, Inc. ARA-TR-98-4854-1

Davis, R. (1998). COMBATING TERRORISM: Observations on Crosscutting Issues.

Testimony Before the Subcommittee on National Security, International Affairs and CriminalJustice, Committee on Government Reform and Oversight, House of Representatives. GAO/T-

NSIAD-98-164

Ettouney, M. and DiMaggio, P. (1998). Integrated Study of Progressive Collapse ofBuildings. Paper from Structural Engineers World Congress. P312-4, ISBN:0-08-042845-2.

Elsevier Science Ltd.

Ettouney, M., Smilowitz, R. and Daddazio, R. (1998). Comparison Between DesignRequirements of Earthquake and Blast Events. Paper from Structural Engineers World

Congress. T210-1, ISBN:0-08-042845-2. Elsevier Science Ltd.

Federal Emergency Management Agency. (1994). NEHRP (National Earthquake Hazards

Reduction Program) Recommended Provisions for Seismic Regulations for New Buildings,Government Printing Office, Washington, DC

Feld, J. and Carper, K. L. (1997). Precast and Prestressed Concrete Structures. ConstructionFailure, John Wiley & Sons, Inc., New York, NY., 9, 302-306

G.A.O. (1998). COMBATING TERRORISM: Threat and Risk Assessments Can HelpPrioritize and Target Program Investments. U.S.G.A.O. Report to Congressional Requesters.

GAO/NSIAD-98-74


42/44

Levy, M. and Salvadori, M. (1992). Big Bangs. Why Buildings Fall Down, W.W. Norton &

Co., New York, NY., 76-83.

National Research Council. (2000). Blast Mitigation for Structures. 1999 Status Report onthe DTRA/TSWG Program. National Academy Press. Washington, D.C.

Sunshine, D. A. (1998). Protection of Structures Subjected to Explosions. Structural

Engineering World Wide 1998. T210-5, ISBN:0-08-042845-2. Elsevier Science Ltd.

Ungar, B. L. (1998). Many Building Security upgrades Made But Problems Have Hindered

Program Implementation. U.S.G.A.O. GSA - Testimony before the Subcommittee on PublicBuildings and Economic Development, Committee on Transportation and Infrastructure, House

of Representatives. GAO/T-GGD-98-141

Wearne, P. (1999). Progressive Collapse, Progressive Security: Oklahoma City and Faulty

Towers: Ronan Point. Collapse: When Buildings Fall Down, TV Books, New York, NY., 117-156.


43/44

APPENDIX A

This appendix provides a copy of shop drawings that were developed by Herrick

corporation during the fabrication of the specimens.

The drawings will be added to

final report after AISC Review iscomplete. A. A.


44/44

APPENDIX B

This Appendix provides plots of all test data channels and information that was collected during

the tests.

The plots are ready and will be

added to final report after AISC

Review is complete. A. A.

Astaneh-AISC-Blast-Resistant-Final Report-Oct 23, 2001a.pdf

Documents

Transcript of Astaneh-AISC-Blast-Resistant-Final Report-Oct 23, 2001a.pdf